Update app.py

app.py

@@ -1,38 +1,37 @@
-#!/usr/bin/env python3
-import os
-import glob
+import aiofiles
+import asyncio
 import base64
-import time
-import shutil
-import zipfile
-import re
+import fitz
+import glob
 import logging
-import asyncio
-import random
-from io import BytesIO
-from datetime import datetime
+import os
+import pandas as pd
 import pytz
-from dataclasses import dataclass
-from typing import Optional
-
+import random
+import re
+import requests
+import shutil
 import streamlit as st
-import pandas as pd
+import time
 import torch
-import fitz
-import requests
-import aiofiles
-from PIL import Image
+import zipfile
+
+from dataclasses import dataclass
+from datetime import datetime
 from diffusers import StableDiffusionPipeline
+from io import BytesIO
+from openai import OpenAI
+from PIL import Image
 from transformers import AutoModelForCausalLM, AutoTokenizer, AutoModel
-from openai import OpenAI  # Updated import for new API
+from typing import Optional
 
-# --- OpenAI Setup ---
+# 🤖 OpenAI wizardry: Summon your API magic!
 client = OpenAI(
     api_key=os.getenv('OPENAI_API_KEY'),
     organization=os.getenv('OPENAI_ORG_ID')
 )
 
-# --- Logging ---
+# 📜 Logging activated: Capturing chaos and calm!
 logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")
 logger = logging.getLogger(__name__)
 log_records = []
@@ -41,7 +40,7 @@ class LogCaptureHandler(logging.Handler):
         log_records.append(record)
 logger.addHandler(LogCaptureHandler())
 
-# --- Streamlit Page Config ---
+# 🎨 Streamlit styling: Designing a cosmic interface!
 st.set_page_config(
     page_title="AI Vision & SFT Titans 🚀",
     page_icon="🤖",
@@ -54,32 +53,20 @@ st.set_page_config(
     }
 )
 
-# --- Session State Defaults ---
-if 'history' not in st.session_state:
-    st.session_state['history'] = []
-if 'builder' not in st.session_state:
-    st.session_state['builder'] = None
-if 'model_loaded' not in st.session_state:
-    st.session_state['model_loaded'] = False
-if 'processing' not in st.session_state:
-    st.session_state['processing'] = {}
-if 'asset_checkboxes' not in st.session_state:
-    st.session_state['asset_checkboxes'] = {}
-if 'downloaded_pdfs' not in st.session_state:
-    st.session_state['downloaded_pdfs'] = {}
-if 'unique_counter' not in st.session_state:
-    st.session_state['unique_counter'] = 0
-if 'selected_model_type' not in st.session_state:
-    st.session_state['selected_model_type'] = "Causal LM"
-if 'selected_model' not in st.session_state:
-    st.session_state['selected_model'] = "None"
-if 'cam0_file' not in st.session_state:
-    st.session_state['cam0_file'] = None
-if 'cam1_file' not in st.session_state:
-    st.session_state['cam1_file'] = None
+st.session_state.setdefault('history', [])  # 🌱 History: starting fresh if empty!
+st.session_state.setdefault('builder', None)  # 🛠️ Builder: set up the builder if it's missing!
+st.session_state.setdefault('model_loaded', False)  # 🚦 Model Loaded: mark as not loaded by default!
+st.session_state.setdefault('processing', {})  # ⏳ Processing: initialize processing state as an empty dict!
+st.session_state.setdefault('asset_checkboxes', {})  # ✅ Asset Checkboxes: default to an empty dictionary!
+st.session_state.setdefault('downloaded_pdfs', {})  # 📄 Downloaded PDFs: start with no PDFs downloaded!
+st.session_state.setdefault('unique_counter', 0)  # 🔢 Unique Counter: initialize the counter to zero!
+st.session_state.setdefault('selected_model_type', "Causal LM")  # 🧠 Selected Model Type: default to "Causal LM"!
+st.session_state.setdefault('selected_model', "None")  # 🤖 Selected Model: set to "None" if not already set!
+st.session_state.setdefault('cam0_file', None)  # 📸 Cam0 File: no file loaded by default!
+st.session_state.setdefault('cam1_file', None)  # 📸 Cam1 File: no file loaded by default!
 
-# --- Model & Diffusion DataClasses ---
-@dataclass
+
+@dataclass  # 🎨 ModelConfig: A blueprint for model configurations!
 class ModelConfig:
     name: str
     base_model: str
@@ -87,44 +74,43 @@ class ModelConfig:
     domain: Optional[str] = None
     model_type: str = "causal_lm"
     @property
-    def model_path(self):
-        return f"models/{self.name}"
+    def model_path(self): return f"models/{self.name}"  # 🚀 Model Path: Home base for brilliance!
 
-@dataclass
+@dataclass  # 🎨 DiffusionConfig: Where diffusion magic takes shape!
 class DiffusionConfig:
     name: str
     base_model: str
     size: str
     domain: Optional[str] = None
     @property
-    def model_path(self):
-        return f"diffusion_models/{self.name}"
-
-# --- Model Builders ---
-class ModelBuilder:
-    def __init__(self):
-        self.config = None
-        self.model = None
-        self.tokenizer = None
-        self.jokes = ["Why did the AI go to therapy? Too many layers to unpack! 😂",
-                      "Training complete! Time for a binary coffee break. ☕"]
-    def load_model(self, model_path: str, config: Optional[ModelConfig] = None):
-        with st.spinner(f"Loading {model_path}... ⏳"):
+    def model_path(self): return f"diffusion_models/{self.name}"  # 🚀 Diffusion Path: Let the diffusion begin!
+
+class ModelBuilder:  # 🔧 ModelBuilder: Crafting AI wonders with wit!
+    def __init__(self):  # 🚀 Initialize: Setting up the AI factory!
+        self.config = None  # No config yet—waiting for genius!
+        self.model = None  # Model not built until the magic happens!
+        self.tokenizer = None  # Tokenizer: Ready to speak in AI!
+        self.jokes = [  # 🤣 Jokes to keep the circuits laughing!
+            "Why did the AI go to therapy? Too many layers to unpack! 😂",
+            "Training complete! Time for a binary coffee break. ☕",
+            "I told my neural network a joke; it couldn't stop dropping bits! 🤖",
+            "I asked the AI for a pun, and it said, 'I'm punning on parallel processing!' 😄",
+            "Debugging my code is like a stand-up routine—always a series of exceptions! 😆"
+        ]
+    def load_model(self, model_path: str, config: Optional[ModelConfig] = None):  # 🔄 load_model: Booting up genius!
+        with st.spinner(f"Loading {model_path}... ⏳"):  # ⏳ Spinner: Genius loading...
             self.model = AutoModelForCausalLM.from_pretrained(model_path)
             self.tokenizer = AutoTokenizer.from_pretrained(model_path)
-            if self.tokenizer.pad_token is None:
-                self.tokenizer.pad_token = self.tokenizer.eos_token
-            if config:
-                self.config = config
-            self.model.to("cuda" if torch.cuda.is_available() else "cpu")
-        st.success(f"Model loaded! 🎉 {random.choice(self.jokes)}")
+            if self.tokenizer.pad_token is None: self.tokenizer.pad_token = self.tokenizer.eos_token  # 🔧 Fix pad token if missing!
+            if config: self.config = config  # 🛠️ Config loaded—setting the stage!
+            self.model.to("cuda" if torch.cuda.is_available() else "cpu")  # 💻 Deploying the model to its device!
+        st.success(f"Model loaded! 🎉 {random.choice(self.jokes)}")  # 🎉 Success: Model is now in orbit!
         return self
-    def save_model(self, path: str):
-        with st.spinner("Saving model... 💾"):
-            os.makedirs(os.path.dirname(path), exist_ok=True)
-            self.model.save_pretrained(path)
-            self.tokenizer.save_pretrained(path)
-        st.success(f"Model saved at {path}! ✅")
+    def save_model(self, path: str):  # 💾 save_model: Securing your masterpiece!
+        with st.spinner("Saving model... 💾"):  # ⏳ Spinner: Saving brilliance...
+            os.makedirs(os.path.dirname(path), exist_ok=True); self.model.save_pretrained(path); self.tokenizer.save_pretrained(path)  # 📂 Directory magic: Creating and saving!
+        st.success(f"Model saved at {path}! ✅")  # ✅ Success: Your model is safely stored!
+
 
 class DiffusionBuilder:
     def __init__(self):
@@ -145,577 +131,387 @@ class DiffusionBuilder:
     def generate(self, prompt: str):
         return self.pipeline(prompt, num_inference_steps=20).images[0]
 
-# --- Utility Functions ---
-def generate_filename(sequence, ext="png"):
-    timestamp = time.strftime("%d%m%Y%H%M%S")
-    return f"{sequence}_{timestamp}.{ext}"
-
-def pdf_url_to_filename(url):
-    safe_name = re.sub(r'[<>:"/\\|?*]', '_', url)
-    return f"{safe_name}.pdf"
-
-def get_download_link(file_path, mime_type="application/pdf", label="Download"):
-    with open(file_path, 'rb') as f:
-        data = f.read()
-    b64 = base64.b64encode(data).decode()
-    return f'<a href="data:{mime_type};base64,{b64}" download="{os.path.basename(file_path)}">{label}</a>'
-
-def zip_directory(directory_path, zip_path):
-    with zipfile.ZipFile(zip_path, 'w', zipfile.ZIP_DEFLATED) as zipf:
-        for root, _, files in os.walk(directory_path):
-            for file in files:
-                zipf.write(os.path.join(root, file), os.path.relpath(os.path.join(root, file), os.path.dirname(directory_path)))
-
-def get_model_files(model_type="causal_lm"):
-    path = "models/*" if model_type == "causal_lm" else "diffusion_models/*"
-    dirs = [d for d in glob.glob(path) if os.path.isdir(d)]
-    return dirs if dirs else ["None"]
-
-def get_gallery_files(file_types=["png", "pdf"]):
-    return sorted(list(set([f for ext in file_types for f in glob.glob(f"*.{ext}")])))  # Deduplicate files
-
-def get_pdf_files():
-    return sorted(glob.glob("*.pdf"))
+def generate_filename(sequence, ext="png"): return f"{sequence}_{time.strftime('%d%m%Y%H%M%S')}.{ext}"  # ⏳ Generate filename with timestamp magic!
+def pdf_url_to_filename(url): return f"{re.sub(r'[<>:\"/\\|?*]', '_', url)}.pdf"  # 📄 Convert URL to a safe PDF filename – no hackers allowed!
+def get_download_link(file_path, mime_type="application/pdf", label="Download"): return f'<a href="data:{mime_type};base64,{base64.b64encode(open(file_path, "rb").read()).decode()}" download="{os.path.basename(file_path)}">{label}</a>'  # 🔗 Create a download link – click it like it's hot!
+def zip_directory(directory_path, zip_path): 
+    with zipfile.ZipFile(zip_path, 'w', zipfile.ZIP_DEFLATED) as zipf: [zipf.write(os.path.join(root, file), os.path.relpath(os.path.join(root, file), os.path.dirname(directory_path))) for root, _, files in os.walk(directory_path) for file in files]  # 🎁 Zip directory: Packing files faster than Santa on Christmas Eve!
+def get_model_files(model_type="causal_lm"): return [d for d in glob.glob("models/*" if model_type == "causal_lm" else "diffusion_models/*") if os.path.isdir(d)] or ["None"]  # 📂 Get model files: Hunting directories like a pro!
+def get_gallery_files(file_types=["png", "pdf"]): return sorted(list({f for ext in file_types for f in glob.glob(f"*.{ext}")}))  # 🖼️ Get gallery files: Finding art in a digital haystack!
+def get_pdf_files(): return sorted(glob.glob("*.pdf"))  # 📄 Get PDF files: Sorted and served – no paper cuts here!
 
+# 📥 Download PDF: Delivering docs faster than a caffeinated courier!
 def download_pdf(url, output_path):
+    try: 
+        response = requests.get(url, stream=True, timeout=10); [open(output_path, "wb").write(chunk) for chunk in response.iter_content(chunk_size=8192)] if response.status_code == 200 else None; ret = True if response.status_code == 200 else False
+    except requests.RequestException as e: 
+        logger.error(f"Failed to download {url}: {e}"); ret = False
+    return ret  
+
+# 📚 Async PDF Snapshot: Snap your PDF pages without blocking—juggle pages like a ninja! 🥷
+async def process_pdf_snapshot(pdf_path, mode="single"):  
+    start_time = time.time(); status = st.empty(); status.text(f"Processing PDF Snapshot ({mode})... (0s)")
     try:
-        response = requests.get(url, stream=True, timeout=10)
-        if response.status_code == 200:
-            with open(output_path, "wb") as f:
-                for chunk in response.iter_content(chunk_size=8192):
-                    f.write(chunk)
-            return True
-    except requests.RequestException as e:
-        logger.error(f"Failed to download {url}: {e}")
-    return False
-
-# --- Original PDF Snapshot & OCR Functions ---
-async def process_pdf_snapshot(pdf_path, mode="single"):
-    start_time = time.time()
-    status = st.empty()
-    status.text(f"Processing PDF Snapshot ({mode})... (0s)")
-    try:
-        doc = fitz.open(pdf_path)
-        output_files = []
-        if mode == "single":
-            page = doc[0]
-            pix = page.get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
-            output_file = generate_filename("single", "png")
-            pix.save(output_file)
-            output_files.append(output_file)
-        elif mode == "twopage":
-            for i in range(min(2, len(doc))):
-                page = doc[i]
-                pix = page.get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
-                output_file = generate_filename(f"twopage_{i}", "png")
-                pix.save(output_file)
-                output_files.append(output_file)
-        elif mode == "allpages":
-            for i in range(len(doc)):
-                page = doc[i]
-                pix = page.get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
-                output_file = generate_filename(f"page_{i}", "png")
-                pix.save(output_file)
-                output_files.append(output_file)
-        doc.close()
-        elapsed = int(time.time() - start_time)
-        status.text(f"PDF Snapshot ({mode}) completed in {elapsed}s!")
-        update_gallery()
-        return output_files
-    except Exception as e:
-        status.error(f"Failed to process PDF: {str(e)}")
-        return []
-
-async def process_ocr(image, output_file):
-    start_time = time.time()
-    status = st.empty()
-    status.text("Processing GOT-OCR2_0... (0s)")
-    tokenizer = AutoTokenizer.from_pretrained("ucaslcl/GOT-OCR2_0", trust_remote_code=True)
-    model = AutoModel.from_pretrained("ucaslcl/GOT-OCR2_0", trust_remote_code=True, torch_dtype=torch.float32).to("cpu").eval()
-    temp_file = f"temp_{int(time.time())}.png"
-    image.save(temp_file)
-    result = model.chat(tokenizer, temp_file, ocr_type='ocr')
-    os.remove(temp_file)
-    elapsed = int(time.time() - start_time)
-    status.text(f"GOT-OCR2_0 completed in {elapsed}s!")
-    async with aiofiles.open(output_file, "w") as f:
-        await f.write(result)
-    update_gallery()
-    return result
-
-async def process_image_gen(prompt, output_file):
-    start_time = time.time()
-    status = st.empty()
-    status.text("Processing Image Gen... (0s)")
-    if st.session_state['builder'] and isinstance(st.session_state['builder'], DiffusionBuilder) and st.session_state['builder'].pipeline:
-        pipeline = st.session_state['builder'].pipeline
-    else:
-        pipeline = StableDiffusionPipeline.from_pretrained("OFA-Sys/small-stable-diffusion-v0", torch_dtype=torch.float32).to("cpu")
-    gen_image = pipeline(prompt, num_inference_steps=20).images[0]
-    elapsed = int(time.time() - start_time)
-    status.text(f"Image Gen completed in {elapsed}s!")
-    gen_image.save(output_file)
-    update_gallery()
-    return gen_image
-
-# --- Updated Function: Process an image (PIL) with a custom prompt using GPT ---
-def process_image_with_prompt(image, prompt, model="gpt-4o-mini", detail="auto"):
-    buffered = BytesIO()
-    image.save(buffered, format="PNG")
-    img_str = base64.b64encode(buffered.getvalue()).decode("utf-8")
-    messages = [{
-        "role": "user",
-        "content": [
-            {"type": "text", "text": prompt},
-            {
-                "type": "image_url",
-                "image_url": {
-                    "url": f"data:image/png;base64,{img_str}",
-                    "detail": detail  # Added detail parameter
-                }
-            }
-        ]
-    }]
-    try:
-        response = client.chat.completions.create(
-            model=model,
-            messages=messages,
-            max_tokens=300
-        )
-        return response.choices[0].message.content
-    except Exception as e:
-        return f"Error processing image with GPT: {str(e)}"
-
-# --- Updated Function: Process text with GPT ---
-def process_text_with_prompt(text, prompt, model="gpt-4o-mini"):
-    messages = [{"role": "user", "content": prompt + "\n\n" + text}]
+        doc = fitz.open(pdf_path); output_files = []
+        if mode == "single": page = doc[0]; pix = page.get_pixmap(matrix=fitz.Matrix(2.0, 2.0)); output_file = generate_filename("single", "png"); pix.save(output_file); output_files.append(output_file)
+        elif mode == "twopage": 
+            for i in range(min(2, len(doc))): page = doc[i]; pix = page.get_pixmap(matrix=fitz.Matrix(2.0, 2.0)); output_file = generate_filename(f"twopage_{i}", "png"); pix.save(output_file); output_files.append(output_file)
+        elif mode == "allpages": 
+            for i in range(len(doc)): page = doc[i]; pix = page.get_pixmap(matrix=fitz.Matrix(2.0, 2.0)); output_file = generate_filename(f"page_{i}", "png"); pix.save(output_file); output_files.append(output_file)
+        doc.close(); elapsed = int(time.time() - start_time); status.text(f"PDF Snapshot ({mode}) completed in {elapsed}s!"); update_gallery(); return output_files
+    except Exception as e: status.error(f"Failed to process PDF: {str(e)}"); return []
+
+# 😎 Async OCR: Convert images to text while your app keeps on groovin'—no blocking, just rocking! 🎸
+async def process_ocr(image, output_file):  
+    start_time = time.time(); status = st.empty(); status.text("Processing GOT-OCR2_0... (0s)")
+    tokenizer = AutoTokenizer.from_pretrained("ucaslcl/GOT-OCR2_0", trust_remote_code=True); model = AutoModel.from_pretrained("ucaslcl/GOT-OCR2_0", trust_remote_code=True, torch_dtype=torch.float32).to("cpu").eval()
+    temp_file = f"temp_{int(time.time())}.png"; image.save(temp_file)
+    result = model.chat(tokenizer, temp_file, ocr_type='ocr'); os.remove(temp_file)
+    elapsed = int(time.time() - start_time); status.text(f"GOT-OCR2_0 completed in {elapsed}s!")
+    async with aiofiles.open(output_file, "w") as f: await f.write(result)
+    update_gallery(); return result
+
+# 🧞 Async Image Gen: Your image genie—wishing up pictures while the event loop keeps the party going! 🎉
+async def process_image_gen(prompt, output_file):  
+    start_time = time.time(); status = st.empty(); status.text("Processing Image Gen... (0s)")
+    pipeline = st.session_state['builder'].pipeline if st.session_state.get('builder') and isinstance(st.session_state['builder'], DiffusionBuilder) and st.session_state['builder'].pipeline else StableDiffusionPipeline.from_pretrained("OFA-Sys/small-stable-diffusion-v0", torch_dtype=torch.float32).to("cpu")
+    gen_image = pipeline(prompt, num_inference_steps=20).images[0]; elapsed = int(time.time() - start_time)
+    status.text(f"Image Gen completed in {elapsed}s!"); gen_image.save(output_file); update_gallery(); return gen_image
+
+# 🖼️ GPT-Image Interpreter: Turning pixels into prose!
+def process_image_with_prompt(image, prompt, model="gpt-4o-mini", detail="auto"): 
+    buffered = BytesIO(); image.save(buffered, format="PNG")  # 💾 Save the image in-memory as PNG—no hard drives harmed!
+    img_str = base64.b64encode(buffered.getvalue()).decode("utf-8")  # 🔐 Encode image data in Base64 for secure, inline transmission!
+    messages = [{"role": "user", "content": [{"type": "text", "text": prompt}, {"type": "image_url", "image_url": {"url": f"data:image/png;base64,{img_str}", "detail": detail}}]}]  # 💬 Build the GPT conversation with your prompt and image!
     try:
-        response = client.chat.completions.create(
-            model=model,
-            messages=messages,
-            max_tokens=300
-        )
-        return response.choices[0].message.content
-    except Exception as e:
-        return f"Error processing text with GPT: {str(e)}"
-
-# --- Sidebar Setup ---
-st.sidebar.subheader("Gallery Settings")
-if 'gallery_size' not in st.session_state:
-    st.session_state['gallery_size'] = 2  # Default value
-st.session_state['gallery_size'] = st.sidebar.slider(
-    "Gallery Size", 
-    1, 10, st.session_state['gallery_size'], 
-    key="gallery_size_slider"
-)
-
-# --- Updated Gallery Function ---
-def update_gallery():
-    all_files = get_gallery_files()
-    if all_files:
-        st.sidebar.subheader("Asset Gallery 📸📖")
-        cols = st.sidebar.columns(2)
-        for idx, file in enumerate(all_files[:st.session_state['gallery_size']]):
-            with cols[idx % 2]:
-                st.session_state['unique_counter'] += 1
-                unique_id = st.session_state['unique_counter']
-                if file.endswith('.png'):
-                    st.image(Image.open(file), caption=os.path.basename(file), use_container_width=True)
-                else:
-                    doc = fitz.open(file)
-                    pix = doc[0].get_pixmap(matrix=fitz.Matrix(0.5, 0.5))
-                    img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
-                    st.image(img, caption=os.path.basename(file), use_container_width=True)
-                    doc.close()
-                checkbox_key = f"asset_{file}_{unique_id}"
-                st.session_state['asset_checkboxes'][file] = st.checkbox(
-                    "Use for SFT/Input", 
-                    value=st.session_state['asset_checkboxes'].get(file, False), 
-                    key=checkbox_key
-                )
-                mime_type = "image/png" if file.endswith('.png') else "application/pdf"
-                st.markdown(get_download_link(file, mime_type, "Snag It! 📥"), unsafe_allow_html=True)
-                if st.button("Zap It! 🗑️", key=f"delete_{file}_{unique_id}"):
-                    os.remove(file)
-                    st.session_state['asset_checkboxes'].pop(file, None)
-                    st.sidebar.success(f"Asset {os.path.basename(file)} vaporized! 💨")
-                    st.rerun()
-
-update_gallery()
+        response = client.chat.completions.create(model=model, messages=messages, max_tokens=300); return response.choices[0].message.content  # 🤖 Invoke GPT’s magic and return its dazzling output!
+    except Exception as e: return f"Error processing image with GPT: {str(e)}"  # ⚠️ Oops—GPT encountered a snag, so we catch and report the error!
+
+# 📝 GPT-Text Alchemist: Merging your prompt and text into digital gold!
+def process_text_with_prompt(text, prompt, model="gpt-4o-mini"):  
+    messages = [{"role": "user", "content": f"{prompt}\n\n{text}"}]  # 🛠️ Constructing the conversation input like a master wordsmith!
+    try: 
+        response = client.chat.completions.create(model=model, messages=messages, max_tokens=300); return response.choices[0].message.content  # 🤖 Summon GPT’s wisdom and return its brilliant answer!
+    except Exception as e: return f"Error processing text with GPT: {str(e)}"  # ⚠️ Oops, GPT stumbled—catching and reporting the error!
+
+st.sidebar.subheader("Gallery Settings")  # 🎨 Sidebar Gallery: Customize your creative space!
+st.session_state.setdefault('gallery_size', 2)  # 🔧 Setting default gallery size to 2 if it's missing!
+st.session_state['gallery_size'] = st.sidebar.slider("Gallery Size", 1, 10, st.session_state['gallery_size'], key="gallery_size_slider")  # 🎚️ Slide to adjust your gallery size and bring balance to your art!
+
+# 📸 Gallery Updater: Making your assets dazzle and disappear faster than a magician's rabbit! 🐇✨
+def update_gallery():  
+    all_files = get_gallery_files()  # 🔍 Grab all gallery files like a digital treasure hunt!
+    if all_files:  # ✅ If assets are found, let the show begin!
+        st.sidebar.subheader("Asset Gallery 📸📖"); cols = st.sidebar.columns(2)  # 🎨 Set up a stylish 2-column layout in the sidebar!
+        for idx, file in enumerate(all_files[:st.session_state['gallery_size']]):  # 🖼️ Loop through your favorite files, limited by gallery size!
+            with cols[idx % 2]:  # 🔄 Alternate columns—because balance is key (and funny)! 
+                st.session_state['unique_counter'] += 1; unique_id = st.session_state['unique_counter']  # 🚀 Increment your asset counter—every asset gets its moment in the spotlight!
+                if file.endswith('.png'): st.image(Image.open(file), caption=os.path.basename(file), use_container_width=True)  # 🖼️ Display the image like a masterpiece!
+                else:  # 📄 For PDFs, we snap their first page like a paparazzo!
+                    doc = fitz.open(file); pix = doc[0].get_pixmap(matrix=fitz.Matrix(0.5, 0.5)); img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples); st.image(img, caption=os.path.basename(file), use_container_width=True); doc.close()
+                checkbox_key = f"asset_{file}_{unique_id}"  # 🔑 Create a unique key—because every asset deserves VIP treatment!
+                st.session_state['asset_checkboxes'][file] = st.checkbox("Use for SFT/Input", value=st.session_state['asset_checkboxes'].get(file, False), key=checkbox_key)  # ✅ Checkbox: Pick your asset for magic (or SFT)!
+                mime_type = "image/png" if file.endswith('.png') else "application/pdf"  # 📎 Determine MIME type—like sorting your socks, but cooler!
+                st.markdown(get_download_link(file, mime_type, "Snag It! 📥"), unsafe_allow_html=True)  # 🔗 Provide a download link—grab your asset faster than a flash sale!
+                if st.button("Zap It! 🗑️", key=f"delete_{file}_{unique_id}"):  # ⚡ "Zap It!" button: Because sometimes you just gotta make stuff disappear!
+                    os.remove(file); st.session_state['asset_checkboxes'].pop(file, None); st.sidebar.success(f"Asset {os.path.basename(file)} vaporized! 💨"); st.rerun()  # 💥 Delete the file and refresh the gallery—poof, it's gone!
+update_gallery()  # 🎉 Launch the gallery update—let the art party commence! (Joke: Why did the asset cross the road? To get zapped on the other side! 😆)
+
+st.sidebar.subheader("Action Logs 📜")  # 📝 Action Logs: Where our system whispers its secrets!
+with st.sidebar: [st.write(f"{record.asctime} - {record.levelname} - {record.message}") for record in log_records]  # 📚 Loop through log records and display them like diary entries!
+
+st.sidebar.subheader("History 📜")  # 🕰️ History: A walk down memory lane, one log at a time!
+with st.sidebar: [st.write(entry) for entry in st.session_state['history']]  # ⏳ Display every historic moment with style!
+
+tabs = st.tabs(["Camera Snap 📷", "Download PDFs 📥", "Test OCR 🔍", "Build Titan 🌱", "Test Image Gen 🎨", "PDF Process 📄", "Image Process 🖼️", "MD Gallery 📚"])  # 🎭 Tabs: Navigate your AI universe like a boss!
+(tab_camera, tab_download, tab_ocr, tab_build, tab_imggen, tab_pdf_process, tab_image_process, tab_md_gallery) = tabs  # 🚀 Unpack the tabs and get ready to explore—because even tabs need to party!
 
-# --- Sidebar Logs & History ---
-st.sidebar.subheader("Action Logs 📜")
-with st.sidebar:
-    for record in log_records:
-        st.write(f"{record.asctime} - {record.levelname} - {record.message}")
-st.sidebar.subheader("History 📜")
-with st.sidebar:
-    for entry in st.session_state['history']:
-        st.write(entry)
-
-# --- Create Tabs ---
-tabs = st.tabs([
-    "Camera Snap 📷",
-    "Download PDFs 📥",
-    "Test OCR 🔍",
-    "Build Titan 🌱",
-    "Test Image Gen 🎨",
-    "PDF Process 📄",
-    "Image Process 🖼️",
-    "MD Gallery 📚"
-])
-(tab_camera, tab_download, tab_ocr, tab_build, tab_imggen, tab_pdf_process, tab_image_process, tab_md_gallery) = tabs
-
-# === Tab: Camera Snap ===
 with tab_camera:
-    st.header("Camera Snap 📷")
-    st.subheader("Single Capture")
-    cols = st.columns(2)
+    st.header("Camera Snap 📷")  # 🎥 Header: Let’s capture those Kodak moments!
+    st.subheader("Single Capture")  # 📸 Subheader: One snap at a time, no double exposure!
+    cols = st.columns(2)  # 🧩 Creating two columns for double-camera action!
+    
     with cols[0]:
-        cam0_img = st.camera_input("Take a picture - Cam 0", key="cam0")
+        cam0_img = st.camera_input("Take a picture - Cam 0", key="cam0")  # 📷 Cam 0: Say cheese!
         if cam0_img:
-            filename = generate_filename("cam0")
-            if st.session_state['cam0_file'] and os.path.exists(st.session_state['cam0_file']):
-                os.remove(st.session_state['cam0_file'])
-            with open(filename, "wb") as f:
-                f.write(cam0_img.getvalue())
-            st.session_state['cam0_file'] = filename
-            entry = f"Snapshot from Cam 0: {filename}"
-            if entry not in st.session_state['history']:
-                st.session_state['history'] = [e for e in st.session_state['history'] if not e.startswith("Snapshot from Cam 0:")] + [entry]
-            st.image(Image.open(filename), caption="Camera 0", use_container_width=True)
-            logger.info(f"Saved snapshot from Camera 0: {filename}")
-            update_gallery()
+            filename = generate_filename("cam0")  # 🏷️ Filename for Cam 0 snapshot generated!
+            if st.session_state['cam0_file'] and os.path.exists(st.session_state['cam0_file']): os.remove(st.session_state['cam0_file'])  # 🗑️ Out with the old Cam 0 snap!
+            with open(filename, "wb") as f: f.write(cam0_img.getvalue())  # 💾 Saving Cam 0 image like a boss!
+            st.session_state['cam0_file'] = filename  # 🔄 Updating session state for Cam 0 file!
+            entry = f"Snapshot from Cam 0: {filename}"  # 📝 History entry: Cam 0 snapshot recorded!
+            if entry not in st.session_state['history']: 
+                st.session_state['history'] = [e for e in st.session_state['history'] if not e.startswith("Snapshot from Cam 0:")] + [entry]  # 🧹 Cleaning and updating history!
+            st.image(Image.open(filename), caption="Camera 0", use_container_width=True)  # 🖼️ Displaying the fresh Cam 0 image!
+            logger.info(f"Saved snapshot from Camera 0: {filename}")  # 🔍 Logging: Cam 0 snapshot saved!
+            update_gallery()  # 🔄 Refreshing gallery to show the new snap!
+    
     with cols[1]:
-        cam1_img = st.camera_input("Take a picture - Cam 1", key="cam1")
+        cam1_img = st.camera_input("Take a picture - Cam 1", key="cam1")  # 📷 Cam 1: Capture your best side!
         if cam1_img:
-            filename = generate_filename("cam1")
-            if st.session_state['cam1_file'] and os.path.exists(st.session_state['cam1_file']):
-                os.remove(st.session_state['cam1_file'])
-            with open(filename, "wb") as f:
-                f.write(cam1_img.getvalue())
-            st.session_state['cam1_file'] = filename
-            entry = f"Snapshot from Cam 1: {filename}"
+            filename = generate_filename("cam1")  # 🏷️ Filename for Cam 1 snapshot generated!
+            if st.session_state['cam1_file'] and os.path.exists(st.session_state['cam1_file']): os.remove(st.session_state['cam1_file'])  # 🗑️ Out with the old Cam 1 snap!
+            with open(filename, "wb") as f: f.write(cam1_img.getvalue())  # 💾 Saving Cam 1 image like a pro!
+            st.session_state['cam1_file'] = filename  # 🔄 Updating session state for Cam 1 file!
+            entry = f"Snapshot from Cam 1: {filename}"  # 📝 History entry: Cam 1 snapshot recorded!
             if entry not in st.session_state['history']:
-                st.session_state['history'] = [e for e in st.session_state['history'] if not e.startswith("Snapshot from Cam 1:")] + [entry]
-            st.image(Image.open(filename), caption="Camera 1", use_container_width=True)
-            logger.info(f"Saved snapshot from Camera 1: {filename}")
-            update_gallery()
+                st.session_state['history'] = [e for e in st.session_state['history'] if not e.startswith("Snapshot from Cam 1:")] + [entry]  # 🧹 Cleaning and updating history!
+            st.image(Image.open(filename), caption="Camera 1", use_container_width=True)  # 🖼️ Displaying the fresh Cam 1 image!
+            logger.info(f"Saved snapshot from Camera 1: {filename}")  # 🔍 Logging: Cam 1 snapshot saved!
+            update_gallery()  # 🔄 Refreshing gallery to show the new snap!
 
 # === Tab: Download PDFs ===
 with tab_download:
-    st.header("Download PDFs 📥")
-    if st.button("Examples 📚"):
-        example_urls = [
-            "https://arxiv.org/pdf/2308.03892",
-            "https://arxiv.org/pdf/1912.01703",
-            "https://arxiv.org/pdf/2408.11039",
-            "https://arxiv.org/pdf/2109.10282",
-            "https://arxiv.org/pdf/2112.10752",
-            "https://arxiv.org/pdf/2308.11236",
-            "https://arxiv.org/pdf/1706.03762",
-            "https://arxiv.org/pdf/2006.11239",
-            "https://arxiv.org/pdf/2305.11207",
-            "https://arxiv.org/pdf/2106.09685",
-            "https://arxiv.org/pdf/2005.11401",
-            "https://arxiv.org/pdf/2106.10504"
-        ]
-        st.session_state['pdf_urls'] = "\n".join(example_urls)
+    st.header("Download PDFs 📥")  # 📥 Header: Ready to snag PDFs like a digital ninja!
+    if st.button("Examples 📚"):  # 📚 Button: Load up some scholarly URLs for instant fun!
+        example_urls = ["https://arxiv.org/pdf/2308.03892", "https://arxiv.org/pdf/1912.01703", "https://arxiv.org/pdf/2408.11039", "https://arxiv.org/pdf/2109.10282", "https://arxiv.org/pdf/2112.10752", "https://arxiv.org/pdf/2308.11236", "https://arxiv.org/pdf/1706.03762", "https://arxiv.org/pdf/2006.11239", "https://arxiv.org/pdf/2305.11207", "https://arxiv.org/pdf/2106.09685", "https://arxiv.org/pdf/2005.11401", "https://arxiv.org/pdf/2106.10504"]; st.session_state['pdf_urls'] = "\n".join(example_urls)  # 📚 Examples loaded into session!
     
-    url_input = st.text_area("Enter PDF URLs (one per line)", value=st.session_state.get('pdf_urls', ""), height=200)
-    if st.button("Robo-Download 🤖"):
-        urls = url_input.strip().split("\n")
-        progress_bar = st.progress(0)
-        status_text = st.empty()
-        total_urls = len(urls)
-        existing_pdfs = get_pdf_files()
+    url_input = st.text_area("Enter PDF URLs (one per line)", value=st.session_state.get('pdf_urls', ""), height=200)  # 📝 Text area: Paste your PDF URLs here—no commas needed!
+    if st.button("Robo-Download 🤖"):  # 🤖 Button: Let Robo-Download flex its digital muscles!
+        urls = url_input.strip().split("\n"); progress_bar = st.progress(0); status_text = st.empty(); total_urls = len(urls); existing_pdfs = get_pdf_files()  # 🚀 Setup: Preparing to download and track progress!
         for idx, url in enumerate(urls):
             if url:
-                output_path = pdf_url_to_filename(url)
-                status_text.text(f"Fetching {idx + 1}/{total_urls}: {os.path.basename(output_path)}...")
+                output_path = pdf_url_to_filename(url); status_text.text(f"Fetching {idx + 1}/{total_urls}: {os.path.basename(output_path)}...")  # 🔍 Fetching PDF: Checking out file name!
                 if output_path not in existing_pdfs:
                     if download_pdf(url, output_path):
-                        st.session_state['downloaded_pdfs'][url] = output_path
-                        logger.info(f"Downloaded PDF from {url} to {output_path}")
-                        entry = f"Downloaded PDF: {output_path}"
-                        if entry not in st.session_state['history']:
-                            st.session_state['history'].append(entry)
-                        st.session_state['asset_checkboxes'][output_path] = True
-                    else:
-                        st.error(f"Failed to nab {url} 😿")
-                else:
-                    st.info(f"Already got {os.path.basename(output_path)}! Skipping... 🐾")
-                    st.session_state['downloaded_pdfs'][url] = output_path
-                progress_bar.progress((idx + 1) / total_urls)
-        status_text.text("Robo-Download complete! 🚀")
-        update_gallery()
-    mode = st.selectbox("Snapshot Mode", ["Single Page (High-Res)", "Two Pages (High-Res)", "All Pages (High-Res)"], key="download_mode")
-    if st.button("Snapshot Selected 📸"):
-        selected_pdfs = [path for path in get_gallery_files() if path.endswith('.pdf') and st.session_state['asset_checkboxes'].get(path, False)]
+                        file_size = os.path.getsize(output_path)  # 📏 File size: Measured in bytes for bragging rights!
+                        st.session_state['downloaded_pdfs'][url] = output_path; logger.info(f"Downloaded PDF from {url} to {output_path}")  # 💾 Download success: File saved and logged!
+                        entry = f"Downloaded PDF: {output_path} ({file_size} bytes)"; 
+                        if entry not in st.session_state['history']: st.session_state['history'].append(entry)  # 📝 History: Recording download details with file size!
+                        st.session_state['asset_checkboxes'][output_path] = True  # ✅ Marking the asset for further magic!
+                    else: st.error(f"Failed to nab {url} 😿")  # ❌ Oops: Download failed, no kitten cuddles here!
+                else: st.info(f"Already got {os.path.basename(output_path)}! Skipping... 🐾"); st.session_state['downloaded_pdfs'][url] = output_path  # 📂 Already downloaded: Avoiding duplicate work!
+                progress_bar.progress((idx + 1) / total_urls)  # 📊 Progress: Moving the download needle forward!
+        status_text.text("Robo-Download complete! 🚀"); update_gallery()  # 🎉 Finished: All PDFs downloaded and gallery refreshed!
+    
+    mode = st.selectbox("Snapshot Mode", ["Single Page (High-Res)", "Two Pages (High-Res)", "All Pages (High-Res)"], key="download_mode")  # 🎛️ Selectbox: Choose your snapshot resolution!
+    if st.button("Snapshot Selected 📸"):  # 📸 Button: Time to snap some PDF snapshots!
+        selected_pdfs = [path for path in get_gallery_files() if path.endswith('.pdf') and st.session_state['asset_checkboxes'].get(path, False)]  # 📄 Filter: Pick only the PDFs marked for snapshotting!
         if selected_pdfs:
             for pdf_path in selected_pdfs:
-                mode_key = {"Single Page (High-Res)": "single", "Two Pages (High-Res)": "twopage", "All Pages (High-Res)": "allpages"}[mode]
-                snapshots = asyncio.run(process_pdf_snapshot(pdf_path, mode_key))
-                for snapshot in snapshots:
-                    st.image(Image.open(snapshot), caption=snapshot, use_container_width=True)
-                    st.session_state['asset_checkboxes'][snapshot] = True
-            update_gallery()
-        else:
-            st.warning("No PDFs selected for snapshotting! Check some boxes in the sidebar.")
+                mode_key = {"Single Page (High-Res)": "single", "Two Pages (High-Res)": "twopage", "All Pages (High-Res)": "allpages"}[mode]; snapshots = asyncio.run(process_pdf_snapshot(pdf_path, mode_key))  # 🔄 Processing: Generate snapshots per selected mode!
+                for snapshot in snapshots: st.image(Image.open(snapshot), caption=snapshot, use_container_width=True); st.session_state['asset_checkboxes'][snapshot] = True  # 🖼️ Display: Show each snapshot and mark it for potential SFT!
+            update_gallery()  # 🔄 Refresh gallery: Let the new snapshots shine!
+        else: st.warning("No PDFs selected for snapshotting! Check some boxes in the sidebar.")  # ⚠️ Warning: No PDFs chosen—select some boxes to make magic happen!
 
 # === Tab: Test OCR ===
 with tab_ocr:
-    st.header("Test OCR 🔍")
-    all_files = get_gallery_files()
+    st.header("Test OCR 🔍")  # 🔍 Header: Time to turn images into text—magic for your eyeballs!
+    all_files = get_gallery_files();  # 📂 Gathering all assets from the gallery!
     if all_files:
-        if st.button("OCR All Assets 🚀"):
-            full_text = "# OCR Results\n\n"
+        if st.button("OCR All Assets 🚀"):  # 🚀 Button: Blast OCR on every asset in one go!
+            full_text = "# OCR Results\n\n";  # 📝 Starting a full OCR report!
             for file in all_files:
-                if file.endswith('.png'):
-                    image = Image.open(file)
-                else:
-                    doc = fitz.open(file)
-                    pix = doc[0].get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
-                    image = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
-                    doc.close()
-                output_file = generate_filename(f"ocr_{os.path.basename(file)}", "txt")
-                result = asyncio.run(process_ocr(image, output_file))
-                full_text += f"## {os.path.basename(file)}\n\n{result}\n\n"
-                entry = f"OCR Test: {file} -> {output_file}"
-                if entry not in st.session_state['history']:
-                    st.session_state['history'].append(entry)
-            md_output_file = f"full_ocr_{int(time.time())}.md"
-            with open(md_output_file, "w") as f:
-                f.write(full_text)
-            st.success(f"Full OCR saved to {md_output_file}")
-            st.markdown(get_download_link(md_output_file, "text/markdown", "Download Full OCR Markdown"), unsafe_allow_html=True)
-        selected_file = st.selectbox("Select Image or PDF", all_files, key="ocr_select")
+                if file.endswith('.png'): image = Image.open(file)  # 🖼️ PNG? Open image directly!
+                else: 
+                    doc = fitz.open(file); pix = doc[0].get_pixmap(matrix=fitz.Matrix(2.0, 2.0)); image = Image.frombytes("RGB", [pix.width, pix.height], pix.samples); doc.close()  # 📄 PDF? Grab a snapshot of the first page!
+                output_file = generate_filename(f"ocr_{os.path.basename(file)}", "txt");  # 💾 Create a unique filename for the OCR text!
+                result = asyncio.run(process_ocr(image, output_file));  # 🤖 Run OCR asynchronously—non-blocking wizardry!
+                full_text += f"## {os.path.basename(file)}\n\n{result}\n\n";  # 📝 Append the OCR result to the full report!
+                entry = f"OCR Test: {file} -> {output_file}";  # 📝 Log this OCR operation!
+                if entry not in st.session_state['history']: st.session_state['history'].append(entry)  # ✅ Update history if this entry is new!
+            md_output_file = f"full_ocr_{int(time.time())}.md";  # 📝 Generate a markdown filename for the full OCR report!
+            with open(md_output_file, "w") as f: f.write(full_text);  # 💾 Write the full OCR report to disk!
+            st.success(f"Full OCR saved to {md_output_file}");  # 🎉 Success: Full OCR report is saved!
+            st.markdown(get_download_link(md_output_file, "text/markdown", "Download Full OCR Markdown"), unsafe_allow_html=True)  # 🔗 Provide a download link for your OCR masterpiece!
+        selected_file = st.selectbox("Select Image or PDF", all_files, key="ocr_select");  # 🔍 Selectbox: Pick an asset for individual OCR!
         if selected_file:
-            if selected_file.endswith('.png'):
-                image = Image.open(selected_file)
-            else:
-                doc = fitz.open(selected_file)
-                pix = doc[0].get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
-                image = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
-                doc.close()
-            st.image(image, caption="Input Image", use_container_width=True)
-            if st.button("Run OCR 🚀", key="ocr_run"):
-                output_file = generate_filename("ocr_output", "txt")
-                st.session_state['processing']['ocr'] = True
-                result = asyncio.run(process_ocr(image, output_file))
-                entry = f"OCR Test: {selected_file} -> {output_file}"
-                if entry not in st.session_state['history']:
-                    st.session_state['history'].append(entry)
-                st.text_area("OCR Result", result, height=200, key="ocr_result")
-                st.success(f"OCR output saved to {output_file}")
-                st.session_state['processing']['ocr'] = False
-            if selected_file.endswith('.pdf') and st.button("OCR All Pages 🚀", key="ocr_all_pages"):
-                doc = fitz.open(selected_file)
-                full_text = f"# OCR Results for {os.path.basename(selected_file)}\n\n"
+            if selected_file.endswith('.png'): image = Image.open(selected_file)  # 🖼️ Open the selected PNG image!
+            else: 
+                doc = fitz.open(selected_file); pix = doc[0].get_pixmap(matrix=fitz.Matrix(2.0, 2.0)); image = Image.frombytes("RGB", [pix.width, pix.height], pix.samples); doc.close()  # 📄 For PDFs, extract a snapshot from the first page!
+            st.image(image, caption="Input Image", use_container_width=True);  # 🖼️ Display the selected asset for OCR review!
+            if st.button("Run OCR 🚀", key="ocr_run"):  # 🚀 Button: Run OCR on the selected asset!
+                output_file = generate_filename("ocr_output", "txt"); st.session_state['processing']['ocr'] = True;  # 💾 Generate output filename and flag processing!
+                result = asyncio.run(process_ocr(image, output_file));  # 🤖 Execute OCR asynchronously!
+                entry = f"OCR Test: {selected_file} -> {output_file}";  # 📝 Create a log entry for this OCR run!
+                if entry not in st.session_state['history']: st.session_state['history'].append(entry);  # ✅ Update history if new!
+                st.text_area("OCR Result", result, height=200, key="ocr_result");  # 📄 Show the OCR result in a text area!
+                st.success(f"OCR output saved to {output_file}"); st.session_state['processing']['ocr'] = False  # 🎉 Success: OCR result saved and processing flag reset!
+            if selected_file.endswith('.pdf') and st.button("OCR All Pages 🚀", key="ocr_all_pages"):  # 📄 Button: Run OCR on every page of a PDF!
+                doc = fitz.open(selected_file); full_text = f"# OCR Results for {os.path.basename(selected_file)}\n\n";  # 📝 Start a report for multi-page PDF OCR!
                 for i in range(len(doc)):
-                    pix = doc[i].get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
-                    image = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
-                    output_file = generate_filename(f"ocr_page_{i}", "txt")
-                    result = asyncio.run(process_ocr(image, output_file))
-                    full_text += f"## Page {i + 1}\n\n{result}\n\n"
-                    entry = f"OCR Test: {selected_file} Page {i + 1} -> {output_file}"
-                    if entry not in st.session_state['history']:
-                        st.session_state['history'].append(entry)
-                md_output_file = f"full_ocr_{os.path.basename(selected_file)}_{int(time.time())}.md"
-                with open(md_output_file, "w") as f:
-                    f.write(full_text)
-                st.success(f"Full OCR saved to {md_output_file}")
-                st.markdown(get_download_link(md_output_file, "text/markdown", "Download Full OCR Markdown"), unsafe_allow_html=True)
+                    pix = doc[i].get_pixmap(matrix=fitz.Matrix(2.0, 2.0)); image = Image.frombytes("RGB", [pix.width, pix.height], pix.samples);  # 🖼️ Capture each page as an image!
+                    output_file = generate_filename(f"ocr_page_{i}", "txt"); result = asyncio.run(process_ocr(image, output_file));  # 💾 Generate filename and process OCR for the page!
+                    full_text += f"## Page {i + 1}\n\n{result}\n\n";  # 📝 Append the page's OCR result to the report!
+                    entry = f"OCR Test: {selected_file} Page {i + 1} -> {output_file}";  # 📝 Log this page's OCR operation!
+                    if entry not in st.session_state['history']: st.session_state['history'].append(entry)  # ✅ Update history if this entry is new!
+                md_output_file = f"full_ocr_{os.path.basename(selected_file)}_{int(time.time())}.md";  # 📝 Create a markdown filename for the full multi-page OCR report!
+                with open(md_output_file, "w") as f: f.write(full_text);  # 💾 Write the full multi-page OCR report to disk!
+                st.success(f"Full OCR saved to {md_output_file}");  # 🎉 Success: Multi-page OCR report is saved!
+                st.markdown(get_download_link(md_output_file, "text/markdown", "Download Full OCR Markdown"), unsafe_allow_html=True)  # 🔗 Provide a download link for the multi-page OCR report!
     else:
-        st.warning("No assets in gallery yet. Use Camera Snap or Download PDFs!")
+        st.warning("No assets in gallery yet. Use Camera Snap or Download PDFs!")  # ⚠️ Warning: Your gallery is empty—capture or download some assets first!
 
 # === Tab: Build Titan ===
 with tab_build:
-    st.header("Build Titan 🌱")
-    model_type = st.selectbox("Model Type", ["Causal LM", "Diffusion"], key="build_type")
-    base_model = st.selectbox("Select Tiny Model", 
-        ["HuggingFaceTB/SmolLM-135M", "Qwen/Qwen1.5-0.5B-Chat"] if model_type == "Causal LM" else 
-        ["OFA-Sys/small-stable-diffusion-v0", "stabilityai/stable-diffusion-2-base"])
-    model_name = st.text_input("Model Name", f"tiny-titan-{int(time.time())}")
-    domain = st.text_input("Target Domain", "general")
-    if st.button("Download Model ⬇️"):
-        config = (ModelConfig if model_type == "Causal LM" else DiffusionConfig)(name=model_name, base_model=base_model, size="small", domain=domain)
-        builder = ModelBuilder() if model_type == "Causal LM" else DiffusionBuilder()
-        builder.load_model(base_model, config)
-        builder.save_model(config.model_path)
-        st.session_state['builder'] = builder
-        st.session_state['model_loaded'] = True
-        st.session_state['selected_model_type'] = model_type
-        st.session_state['selected_model'] = config.model_path
-        entry = f"Built {model_type} model: {model_name}"
-        if entry not in st.session_state['history']:
-            st.session_state['history'].append(entry)
-        st.success(f"Model downloaded and saved to {config.model_path}! 🎉")
-        st.rerun()
+    st.header("Build Titan 🌱")  # 🌱 Header: Build your own Titan—tiny models, huge ambitions!
+    model_type = st.selectbox("Model Type", ["Causal LM", "Diffusion"], key="build_type")  # 🔍 Choose your model flavor!
+    base_model = st.selectbox(
+        "Select Tiny Model",
+        ["HuggingFaceTB/SmolLM-135M", "Qwen/Qwen1.5-0.5B-Chat"] if model_type == "Causal LM" 
+        else ["OFA-Sys/small-stable-diffusion-v0", "stabilityai/stable-diffusion-2-base"]
+    )  # 🤖 Pick a tiny model based on your choice!
+    model_name = st.text_input("Model Name", f"tiny-titan-{int(time.time())}")  # 🏷️ Auto-generate a cool model name with a timestamp!
+    domain = st.text_input("Target Domain", "general")  # 🎯 Specify your target domain (default: general)!
+    if st.button("Download Model ⬇️"):  # ⬇️ Button: Download your model and get ready to unleash the Titan!
+        config = (ModelConfig if model_type == "Causal LM" else DiffusionConfig)(
+            name=model_name, base_model=base_model, size="small", domain=domain
+        )  # 📝 Create model configuration on the fly!
+        builder = ModelBuilder() if model_type == "Causal LM" else DiffusionBuilder()  # 🔧 Instantiate the builder for your model type!
+        builder.load_model(base_model, config); builder.save_model(config.model_path)  # 🚀 Load and save the model—instant Titan assembly!
+        st.session_state['builder'] = builder; st.session_state['model_loaded'] = True  # ⚙️ Update session state: model is now loaded!
+        st.session_state['selected_model_type'] = model_type; st.session_state['selected_model'] = config.model_path  # 🔑 Store your selection for posterity!
+        entry = f"Built {model_type} model: {model_name}"  # 📝 Log the build event in history!
+        if entry not in st.session_state['history']: st.session_state['history'].append(entry)
+        st.success(f"Model downloaded and saved to {config.model_path}! 🎉"); st.rerun()  # 🎉 Success: Titan built, now re-run to refresh the interface!
 
 # === Tab: Test Image Gen ===
 with tab_imggen:
-    st.header("Test Image Gen 🎨")
-    all_files = get_gallery_files()
+    st.header("Test Image Gen 🎨")  # 🎨 Header: Time to get creative with AI image generation!
+    all_files = get_gallery_files()  # 📂 Retrieve all gallery assets for selection.
     if all_files:
-        selected_file = st.selectbox("Select Image or PDF", all_files, key="gen_select")
+        selected_file = st.selectbox("Select Image or PDF", all_files, key="gen_select")  # 🔍 Select an asset to spark creativity!
         if selected_file:
-            if selected_file.endswith('.png'):
-                image = Image.open(selected_file)
+            if selected_file.endswith('.png'): 
+                image = Image.open(selected_file)  # 🖼️ Directly open PNG images!
             else:
-                doc = fitz.open(selected_file)
-                pix = doc[0].get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
-                image = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
-                doc.close()
-            st.image(image, caption="Reference Image", use_container_width=True)
-            prompt = st.text_area("Prompt", "Generate a neon superhero version of this image", key="gen_prompt")
-            if st.button("Run Image Gen 🚀", key="gen_run"):
-                output_file = generate_filename("gen_output", "png")
-                st.session_state['processing']['gen'] = True
-                result = asyncio.run(process_image_gen(prompt, output_file))
-                entry = f"Image Gen Test: {prompt} -> {output_file}"
-                if entry not in st.session_state['history']:
-                    st.session_state['history'].append(entry)
-                st.image(result, caption="Generated Image", use_container_width=True)
-                st.success(f"Image saved to {output_file}")
-                st.session_state['processing']['gen'] = False
+                doc = fitz.open(selected_file); pix = doc[0].get_pixmap(matrix=fitz.Matrix(2.0, 2.0)); 
+                image = Image.frombytes("RGB", [pix.width, pix.height], pix.samples); doc.close()  # 📄 For PDFs, extract the first page as an image!
+            st.image(image, caption="Reference Image", use_container_width=True)  # 🖼️ Display the chosen asset as reference.
+            prompt = st.text_area("Prompt", "Generate a neon superhero version of this image", key="gen_prompt")  # ✍️ Enter a creative prompt to transform the image!
+            if st.button("Run Image Gen 🚀", key="gen_run"):  # 🚀 Button: Ignite the image generator!
+                output_file = generate_filename("gen_output", "png"); st.session_state['processing']['gen'] = True  # 💾 Create output filename and flag processing status.
+                result = asyncio.run(process_image_gen(prompt, output_file))  # 🤖 Run the async image generation—non-blocking magic in action!
+                entry = f"Image Gen Test: {prompt} -> {output_file}"  # 📝 Log the image generation event!
+                if entry not in st.session_state['history']: st.session_state['history'].append(entry)
+                st.image(result, caption="Generated Image", use_container_width=True)  # 🖼️ Showcase the newly generated image!
+                st.success(f"Image saved to {output_file}"); st.session_state['processing']['gen'] = False  # 🎉 Success: Your masterpiece is saved and processing is complete!
     else:
-        st.warning("No images or PDFs in gallery yet. Use Camera Snap or Download PDFs!")
-    update_gallery()
+        st.warning("No images or PDFs in gallery yet. Use Camera Snap or Download PDFs!")  # ⚠️ Warning: No assets available—capture or download some first!
+    update_gallery()  # 🔄 Refresh the gallery to display any updates!
 
 # === Updated Tab: PDF Process ===
 with tab_pdf_process:
-    st.header("PDF Process")
-    st.subheader("Upload PDFs for GPT-based text extraction")
-    gpt_models = ["gpt-4o", "gpt-4o-mini"]  # Add more vision-capable models as needed
-    selected_gpt_model = st.selectbox("Select GPT Model", gpt_models, key="pdf_gpt_model")
-    detail_level = st.selectbox("Detail Level", ["auto", "low", "high"], key="pdf_detail_level")
-    uploaded_pdfs = st.file_uploader("Upload PDF files", type=["pdf"], accept_multiple_files=True, key="pdf_process_uploader")
-    view_mode = st.selectbox("View Mode", ["Single Page", "Double Page"], key="pdf_view_mode")
-    if st.button("Process Uploaded PDFs", key="process_pdfs"):
-        combined_text = ""
-        for pdf_file in uploaded_pdfs:
-            pdf_bytes = pdf_file.read()
-            temp_pdf_path = f"temp_{pdf_file.name}"
-            with open(temp_pdf_path, "wb") as f:
-                f.write(pdf_bytes)
+    st.header("PDF Process")  # 📄 Header: Ready to transform your PDFs into text with GPT magic!
+    st.subheader("Upload PDFs for GPT-based text extraction")  # 🚀 Subheader: Upload your PDFs and let the AI do the reading!
+    gpt_models = ["gpt-4o", "gpt-4o-mini"]  # 🤖 GPT Models: Pick your AI wizard—more vision-capable models may join the party!
+    selected_gpt_model = st.selectbox("Select GPT Model", gpt_models, key="pdf_gpt_model")  # 🔍 Select your GPT model and let it work its charm!
+    detail_level = st.selectbox("Detail Level", ["auto", "low", "high"], key="pdf_detail_level")  # 🎚️ Detail Level: Fine-tune your extraction’s precision!
+    uploaded_pdfs = st.file_uploader("Upload PDF files", type=["pdf"], accept_multiple_files=True, key="pdf_process_uploader")  # 📤 Uploader: Drag & drop your PDFs for processing!
+    view_mode = st.selectbox("View Mode", ["Single Page", "Double Page"], key="pdf_view_mode")  # 👀 View Mode: Choose single or double page snapshots!
+    
+    if st.button("Process Uploaded PDFs", key="process_pdfs"):  # ⚙️ Button: Kick off the PDF processing extravaganza!
+        combined_text = ""  # 📝 Initialize a blank slate for the GPT output!
+        for pdf_file in uploaded_pdfs:  # 🔄 Loop through each uploaded PDF file!
+            pdf_bytes = pdf_file.read()  # 📥 Read the PDF bytes into memory!
+            temp_pdf_path = f"temp_{pdf_file.name}"  # 🏷️ Create a temporary filename for processing!
+            with open(temp_pdf_path, "wb") as f: f.write(pdf_bytes)  # 💾 Write the PDF to a temporary file!
             try:
-                doc = fitz.open(temp_pdf_path)
-                st.write(f"Processing {pdf_file.name} with {len(doc)} pages")
-                if view_mode == "Single Page":
+                doc = fitz.open(temp_pdf_path)  # 📄 Open the temporary PDF document!
+                st.write(f"Processing {pdf_file.name} with {len(doc)} pages")  # 🔍 Log: Display file name and page count!
+                if view_mode == "Single Page":  # 📑 Single Page Mode: Process each page separately!
                     for i, page in enumerate(doc):
-                        pix = page.get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
-                        img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
-                        st.image(img, caption=f"{pdf_file.name} Page {i+1}")
-                        gpt_text = process_image_with_prompt(img, "Extract the electronic text from image", model=selected_gpt_model, detail=detail_level)
-                        combined_text += f"\n## {pdf_file.name} - Page {i+1}\n\n{gpt_text}\n"
-                else:  # Double Page
-                    pages = list(doc)
+                        pix = page.get_pixmap(matrix=fitz.Matrix(2.0, 2.0));  # 🎞️ Create a high-res pixmap of the page!
+                        img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples);  # 🖼️ Convert the pixmap to an image!
+                        st.image(img, caption=f"{pdf_file.name} Page {i+1}");  # 🖼️ Display the page image!
+                        gpt_text = process_image_with_prompt(
+                            img, "Extract the electronic text from image", model=selected_gpt_model, detail=detail_level
+                        );  # 🤖 Run GPT to extract text from the image!
+                        combined_text += f"\n## {pdf_file.name} - Page {i+1}\n\n{gpt_text}\n";  # 📝 Append the result to the combined text!
+                else:  # 📄 Double Page Mode: Process pages in pairs!
+                    pages = list(doc);  # 🔢 Convert document pages to a list!
                     for i in range(0, len(pages), 2):
-                        if i+1 < len(pages):
-                            pix1 = pages[i].get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
-                            img1 = Image.frombytes("RGB", [pix1.width, pix1.height], pix1.samples)
-                            pix2 = pages[i+1].get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
-                            img2 = Image.frombytes("RGB", [pix2.width, pix2.height], pix2.samples)
-                            total_width = img1.width + img2.width
-                            max_height = max(img1.height, img2.height)
-                            combined_img = Image.new("RGB", (total_width, max_height))
-                            combined_img.paste(img1, (0, 0))
-                            combined_img.paste(img2, (img1.width, 0))
-                            st.image(combined_img, caption=f"{pdf_file.name} Pages {i+1}-{i+2}")
-                            gpt_text = process_image_with_prompt(combined_img, "Extract the electronic text from image", model=selected_gpt_model, detail=detail_level)
-                            combined_text += f"\n## {pdf_file.name} - Pages {i+1}-{i+2}\n\n{gpt_text}\n"
-                        else:
-                            pix = pages[i].get_pixmap(matrix=fitz.Matrix(2.0, 2.0))
-                            img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
-                            st.image(img, caption=f"{pdf_file.name} Page {i+1}")
-                            gpt_text = process_image_with_prompt(img, "Extract the electronic text from image", model=selected_gpt_model, detail=detail_level)
-                            combined_text += f"\n## {pdf_file.name} - Page {i+1}\n\n{gpt_text}\n"
-                doc.close()
-            except Exception as e:
-                st.error(f"Error processing {pdf_file.name}: {str(e)}")
-            finally:
-                os.remove(temp_pdf_path)
-        output_filename = generate_filename("processed_pdf", "md")
-        with open(output_filename, "w", encoding="utf-8") as f:
-            f.write(combined_text)
-        st.success(f"PDF processing complete. MD file saved as {output_filename}")
-        st.markdown(get_download_link(output_filename, "text/markdown", "Download Processed PDF MD"), unsafe_allow_html=True)
+                        if i+1 < len(pages):  # 👯 Process two pages if available!
+                            pix1 = pages[i].get_pixmap(matrix=fitz.Matrix(2.0, 2.0)); img1 = Image.frombytes("RGB", [pix1.width, pix1.height], pix1.samples);  # 🖼️ Process first page!
+                            pix2 = pages[i+1].get_pixmap(matrix=fitz.Matrix(2.0, 2.0)); img2 = Image.frombytes("RGB", [pix2.width, pix2.height], pix2.samples);  # 🖼️ Process second page!
+                            total_width = img1.width + img2.width; max_height = max(img1.height, img2.height);  # 📏 Calculate dimensions for the combined image!
+                            combined_img = Image.new("RGB", (total_width, max_height));  # 🖼️ Create a blank canvas for the two pages!
+                            combined_img.paste(img1, (0, 0)); combined_img.paste(img2, (img1.width, 0));  # 🎨 Paste the images side by side!
+                            st.image(combined_img, caption=f"{pdf_file.name} Pages {i+1}-{i+2}");  # 🖼️ Display the combined image!
+                            gpt_text = process_image_with_prompt(
+                                combined_img, "Extract the electronic text from image", model=selected_gpt_model, detail=detail_level
+                            );  # 🤖 Extract text from the combined image!
+                            combined_text += f"\n## {pdf_file.name} - Pages {i+1}-{i+2}\n\n{gpt_text}\n";  # 📝 Append the result to the combined text!
+                        else:  # 🔹 If there's an odd page out, process it solo!
+                            pix = pages[i].get_pixmap(matrix=fitz.Matrix(2.0, 2.0)); img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples);  # 🖼️ Process the single remaining page!
+                            st.image(img, caption=f"{pdf_file.name} Page {i+1}");  # 🖼️ Display the solo page image!
+                            gpt_text = process_image_with_prompt(
+                                img, "Extract the electronic text from image", model=selected_gpt_model, detail=detail_level
+                            );  # 🤖 Run GPT extraction on the solo page!
+                            combined_text += f"\n## {pdf_file.name} - Page {i+1}\n\n{gpt_text}\n";  # 📝 Append the result!
+                doc.close();  # ✅ Close the PDF document to free up resources!
+            except Exception as e: 
+                st.error(f"Error processing {pdf_file.name}: {str(e)}");  # ⚠️ Error: Report any issues during processing!
+            finally: 
+                os.remove(temp_pdf_path);  # 🧹 Cleanup: Remove the temporary PDF file!
+        output_filename = generate_filename("processed_pdf", "md");  # 🏷️ Generate a unique filename for the Markdown output!
+        with open(output_filename, "w", encoding="utf-8") as f: f.write(combined_text);  # 💾 Write the combined GPT text to the Markdown file!
+        st.success(f"PDF processing complete. MD file saved as {output_filename}");  # 🎉 Success: Notify the user of completion!
+        st.markdown(get_download_link(output_filename, "text/markdown", "Download Processed PDF MD"), unsafe_allow_html=True);  # 🔗 Provide a download link for your processed file!
 
 # === Updated Tab: Image Process ===
 with tab_image_process:
-    st.header("Image Process")
-    st.subheader("Upload Images for GPT-based OCR")
-    gpt_models = ["gpt-4o", "gpt-4o-mini"]  # Add more vision-capable models as needed
-    selected_gpt_model = st.selectbox("Select GPT Model", gpt_models, key="img_gpt_model")
-    detail_level = st.selectbox("Detail Level", ["auto", "low", "high"], key="img_detail_level")
-    prompt_img = st.text_input("Enter prompt for image processing", "Extract the electronic text from image", key="img_process_prompt")
-    uploaded_images = st.file_uploader("Upload image files", type=["png", "jpg", "jpeg"], accept_multiple_files=True, key="image_process_uploader")
-    if st.button("Process Uploaded Images", key="process_images"):
-        combined_text = ""
+    st.header("Image Process")  # 🖼️ Header: Transform images into text with GPT magic!
+    st.subheader("Upload Images for GPT-based OCR")  # 🚀 Subheader: Let your images speak for themselves!
+    gpt_models = ["gpt-4o", "gpt-4o-mini"]  # 🤖 GPT Models: Choose your image wizard!
+    selected_gpt_model = st.selectbox("Select GPT Model", gpt_models, key="img_gpt_model")  # 🔍 Pick your GPT model for image processing!
+    detail_level = st.selectbox("Detail Level", ["auto", "low", "high"], key="img_detail_level")  # 🎚️ Detail Level: Set your extraction precision!
+    prompt_img = st.text_input("Enter prompt for image processing", "Extract the electronic text from image", key="img_process_prompt")  # ✍️ Prompt: Tell GPT what to extract!
+    uploaded_images = st.file_uploader("Upload image files", type=["png", "jpg", "jpeg"], accept_multiple_files=True, key="image_process_uploader")  # 📤 Uploader: Drag & drop your images here!
+    if st.button("Process Uploaded Images", key="process_images"):  # 🚀 Button: Fire up the image processing!
+        combined_text = ""  # 📝 Initialize combined text output!
         for img_file in uploaded_images:
             try:
-                img = Image.open(img_file)
-                st.image(img, caption=img_file.name)
-                gpt_text = process_image_with_prompt(img, prompt_img, model=selected_gpt_model, detail=detail_level)
-                combined_text += f"\n## {img_file.name}\n\n{gpt_text}\n"
-            except Exception as e:
-                st.error(f"Error processing image {img_file.name}: {str(e)}")
-        output_filename = generate_filename("processed_image", "md")
-        with open(output_filename, "w", encoding="utf-8") as f:
-            f.write(combined_text)
-        st.success(f"Image processing complete. MD file saved as {output_filename}")
-        st.markdown(get_download_link(output_filename, "text/markdown", "Download Processed Image MD"), unsafe_allow_html=True)
+                img = Image.open(img_file); st.image(img, caption=img_file.name)  # 📸 Display each uploaded image!
+                gpt_text = process_image_with_prompt(img, prompt_img, model=selected_gpt_model, detail=detail_level)  # 🤖 Process image with GPT magic!
+                combined_text += f"\n## {img_file.name}\n\n{gpt_text}\n"  # 📝 Append GPT output with file header!
+            except Exception as e: st.error(f"Error processing image {img_file.name}: {str(e)}")  # ⚠️ Oops: Report errors if any!
+        output_filename = generate_filename("processed_image", "md")  # 💾 Generate a unique filename for the Markdown output!
+        with open(output_filename, "w", encoding="utf-8") as f: f.write(combined_text)  # 📝 Save the combined GPT output!
+        st.success(f"Image processing complete. MD file saved as {output_filename}")  # 🎉 Success: Notify the user!
+        st.markdown(get_download_link(output_filename, "text/markdown", "Download Processed Image MD"), unsafe_allow_html=True)  # 🔗 Provide a download link!
 
 # === Updated Tab: MD Gallery ===
 with tab_md_gallery:
-    st.header("MD Gallery and GPT Processing")
-    gpt_models = ["gpt-4o", "gpt-4o-mini"]  # Add more vision-capable models as needed
-    selected_gpt_model = st.selectbox("Select GPT Model", gpt_models, key="md_gpt_model")
-    md_files = sorted(glob.glob("*.md"))
+    st.header("MD Gallery and GPT Processing")  # 📚 Header: Where markdown meets GPT wizardry!
+    gpt_models = ["gpt-4o", "gpt-4o-mini"]  # 🤖 GPT Models: Pick your processing partner!
+    selected_gpt_model = st.selectbox("Select GPT Model", gpt_models, key="md_gpt_model")  # 🔍 Select a GPT model for MD processing!
+    md_files = sorted(glob.glob("*.md"))  # 📂 Gather all Markdown files in the directory!
     if md_files:
-        st.subheader("Individual File Processing")
-        cols = st.columns(2)
+        st.subheader("Individual File Processing")  # 🔍 Subheader: Process files one at a time!
+        cols = st.columns(2)  # 🧩 Set up two columns for a balanced view!
         for idx, md_file in enumerate(md_files):
             with cols[idx % 2]:
-                st.write(md_file)
-                if st.button(f"Process {md_file}", key=f"process_md_{md_file}"):
+                st.write(md_file)  # 📄 Show the filename!
+                if st.button(f"Process {md_file}", key=f"process_md_{md_file}"):  # 🚀 Button: Process this file!
                     try:
-                        with open(md_file, "r", encoding="utf-8") as f:
-                            content = f.read()
-                        prompt_md = "Summarize this into markdown outline with emojis and number the topics 1..12"
-                        result_text = process_text_with_prompt(content, prompt_md, model=selected_gpt_model)
-                        st.markdown(result_text)
-                        output_filename = generate_filename(f"processed_{os.path.splitext(md_file)[0]}", "md")
-                        with open(output_filename, "w", encoding="utf-8") as f:
-                            f.write(result_text)
-                        st.markdown(get_download_link(output_filename, "text/markdown", f"Download {output_filename}"), unsafe_allow_html=True)
-                    except Exception as e:
-                        st.error(f"Error processing {md_file}: {str(e)}")
-        st.subheader("Batch Processing")
-        st.write("Select MD files to combine and process:")
-        selected_md = {}
-        for md_file in md_files:
-            selected_md[md_file] = st.checkbox(md_file, key=f"checkbox_md_{md_file}")
-        batch_prompt = st.text_input("Enter batch processing prompt", "Summarize this into markdown outline with emojis and number the topics 1..12", key="batch_prompt")
-        if st.button("Process Selected MD Files", key="process_batch_md"):
-            combined_content = ""
+                        with open(md_file, "r", encoding="utf-8") as f: content = f.read()  # 📖 Read file content!
+                        prompt_md = "Summarize this into markdown outline with emojis and number the topics 1..12"  # ✍️ Prompt: Summarize with style!
+                        result_text = process_text_with_prompt(content, prompt_md, model=selected_gpt_model)  # 🤖 Let GPT work its magic!
+                        st.markdown(result_text)  # 🎨 Display the GPT output!
+                        output_filename = generate_filename(f"processed_{os.path.splitext(md_file)[0]}", "md")  # 💾 Create a unique output filename!
+                        with open(output_filename, "w", encoding="utf-8") as f: f.write(result_text)  # 📝 Save the processed content!
+                        st.markdown(get_download_link(output_filename, "text/markdown", f"Download {output_filename}"), unsafe_allow_html=True)  # 🔗 Provide a download link!
+                    except Exception as e: st.error(f"Error processing {md_file}: {str(e)}")  # ⚠️ Report errors if processing fails!
+        st.subheader("Batch Processing")  # 📚 Subheader: Combine and process multiple files at once!
+        st.write("Select MD files to combine and process:")  # 🔍 Instruction: Choose files for batch processing!
+        selected_md = {}  # 🗂️ Initialize selection dictionary!
+        for md_file in md_files: selected_md[md_file] = st.checkbox(md_file, key=f"checkbox_md_{md_file}")  # ✅ Create checkboxes for each file!
+        batch_prompt = st.text_input("Enter batch processing prompt", "Summarize this into markdown outline with emojis and number the topics 1..12", key="batch_prompt")  # ✍️ Batch prompt: Set your summarization style!
+        if st.button("Process Selected MD Files", key="process_batch_md"):  # 🚀 Button: Process the selected files!
+            combined_content = ""  # 📝 Initialize combined content string!
             for md_file, selected in selected_md.items():
                 if selected:
                     try:
-                        with open(md_file, "r", encoding="utf-8") as f:
-                            combined_content += f"\n## {md_file}\n" + f.read() + "\n"
-                    except Exception as e:
-                        st.error(f"Error reading {md_file}: {str(e)}")
+                        with open(md_file, "r", encoding="utf-8") as f: combined_content += f"\n## {md_file}\n" + f.read() + "\n"  # 📄 Append each selected file's content!
+                    except Exception as e: st.error(f"Error reading {md_file}: {str(e)}")  # ⚠️ Report errors if file reading fails!
             if combined_content:
-                result_text = process_text_with_prompt(combined_content, batch_prompt, model=selected_gpt_model)
-                st.markdown(result_text)
-                output_filename = generate_filename("batch_processed_md", "md")
-                with open(output_filename, "w", encoding="utf-8") as f:
-                    f.write(result_text)
-                st.success(f"Batch processing complete. MD file saved as {output_filename}")
-                st.markdown(get_download_link(output_filename, "text/markdown", "Download Batch Processed MD"), unsafe_allow_html=True)
+                result_text = process_text_with_prompt(combined_content, batch_prompt, model=selected_gpt_model)  # 🤖 Process the batch with GPT!
+                st.markdown(result_text)  # 🎨 Display the combined GPT output!
+                output_filename = generate_filename("batch_processed_md", "md")  # 💾 Generate a unique filename for the batch output!
+                with open(output_filename, "w", encoding="utf-8") as f: f.write(result_text)  # 📝 Save the batch processed text!
+                st.success(f"Batch processing complete. MD file saved as {output_filename}")  # 🎉 Notify success!
+                st.markdown(get_download_link(output_filename, "text/markdown", "Download Batch Processed MD"), unsafe_allow_html=True)  # 🔗 Provide a download link!
             else:
-                st.warning("No MD files selected.")
+                st.warning("No MD files selected.")  # ⚠️ Warning: No files were chosen for batch processing!
     else:
-        st.warning("No MD files found.")
+        st.warning("No MD files found.")  # ⚠️ Warning: Your gallery is empty—no markdown files available!